“Talking to computers” has been a staple of science fiction for decades. Conversations between humans and computers requires at its core the ability to understand human language—“natural language processing” (NLP). NLP analyzes text—either typed text or text created by speech recognition technology.
Today’s NLP works best within a limited context. If a customer calls a bank, the NLP in an automated system is limited to the banking context; it needs to understand such requests as how to transfer money between accounts. If the customer calls for pizza delivery, the vocabulary and intent of the user is completely different.
Given a sufficiently limited context, today’s NLP works well (presuming it has been developed properly). The general digital assistants, e.g., Amazon’s Alexa, Google Assistant, and Apple’s Siri, illustrate the huge scope of intents possible, but the challenge is much less for a context limited to a specific company or application. This eases the development of company- or application-specific chatbots or digital assistants.
What made the advance in NLP and speech recognition possible? Some invention was required, but fundamentally the breakthrough was simply computer power crossing a threshold that allowed it to analyze huge amounts of labelled text and speech and model the connections between the labels and the speech or text—most powerfully by using machine learning and deep-neural-network models. In addition to making the analysis of very large databases economically feasible, computer power passed a threshold that allowed the resulting complex models to be computed in “real-time,” fast enough to support an interactive conversation.
NLP progress motivated development of supporting technology, including text-to-speech synthesis to speak back to a user and dialog management that keeps track of the full context of a conversation. Other general IT trends supported the effective use of conversational technology, e.g., Application Programming Interfaces to enterprise software allows chatbots and digital assistants to provide answers from company databases and to personalize conversations.
NLP and speech recognition have already become a major innovation in user interfaces, with the general digital assistants becoming an increasingly powerful alternative to web search. Conversational functionality will come to be expected in most computer applications as a major supplement to—if not a replacement for—the increasingly over-burdened point-and-click Graphical User Interface. We are so accustomed to the GUI that we forget how it revolutionized personal computing and mobile phones. Conversational technology will have a similar impact on how we use digital systems.
A company can take early advantage of a major trend or play catch-up later. You wouldn’t think of not having a company web site today. It will soon be unthinkable not to have a company chatbot and/or digital assistant.
Automating customer service using conversational technology is one of the most advanced applications providing major savings and improvements in the customer experience. A call to a customer service line can be answered by a prompt such as “Please say why you are calling” with an automated system either providing customers what they are seeking or transferring them to an appropriate agent. A digital assistant or chatbot provides your customer with quick results—not a series of frustrating options or a long wait for an agent.
This opportunity may also become a competitive necessity as the general personal assistants, such as Google Assistant, Amazon’s Alexa, and Apple’s Siri increasingly become like web portals, not only providing information directly, but connecting users to outside services and outside-party customer support. Companies will soon find they are expected to have a company digital assistant available through these voice portals, just as they are expected to have a web site available through a web browser.
Whether you contact your customers or they contact you, a well-done company digital assistant can expand your connection with your customer. Instead of getting rid of a customer as quickly as possible, an automated system makes it highly cost-effective to expand the customer connection to market additional products or services. The wall between customer service and marketing can crumble.
And company employees can use a digital assistant to quickly get answers to human resources questions or connect with enterprise software more efficiently. The first solution increases the efficiency of supporting HR requests and the second can make enterprise software more fully and accurately exploited. Supporting the sales process is one successful application, allowing, for example, sales personnel to update sales databases or access customer information even while driving.
Beyond automating conversations, NLP and speech recognition also allow effective analysis of unstructured text or voice data, e.g., recorded call center calls. Analytics software can use these technologies to find patterns in this data, e.g., a large number of calls reporting an issue not previously identified.
The characteristics that let a human recognize a voice have also been automated as “speaker identification” or “voice biometrics.” This technology can be used for security or to identify who is speaking when there are multiple speakers in a conversation.
Few companies are equipped to employ advanced conversational technology without a learning curve or help. Fortunately, there are many vendors with tools, even template-like full solutions, that can support such efforts. The services range from building a full digital assistant customized to a company’s needs to providing parts of a solution, such as speech recognition technology, that can be combined by a company’s development team into a solution.
The Bots & Assistants Conference is designed to be a major resource that will highlight these tools and services, as well as case studies that show what works and what to avoid. Beyond talks by knowledgeable experts, the demo derby will let you view examples presented by the companies themselves in brief videos in any number or order you prefer.
Small breakout sessions after the talks will allow you to hear and express individual opinions and experiences. Some speakers will participate, allowing a longer question-and-answer session if desired.